Ubiquitous Computing and Communication Journal

EFFICIENT IMPLEMENTATION OF DOWNLINK CDMA

EQUALIZATION USING FREQUENCY DOMAIN
APPROXIMATION
F. S. Al-kamali+, M. I. Dessouky, B. M. Sallam, and F. E. El-Samie++
Department of Electrical Communications
Faculty of Electronic Engineering
Menoufia University
Menouf, Egypt
E-mails: +faisalalkamali@yahoo.com, ++fathi_sayed@yahoo.com

ABSTRACT
A signal transimitted through a wireless channel may be severely distorted due to
intersymbol interference (ISI) and multiple access interference (MAI). In this paper,
we propose an efficient CDMA receiver based on a frequency domain equalization
with a regularized zero forcing equalizer and unit clipper decision with parallel
interference cancellation (FDE-RZF-CPIC) to combat both ISI and MAI. This
receiver is suitable for downlink zero padding CDMA (ZP-CDMA) cellular systems.
The effects of the decision function, the channel estimation, the number of cancelled
users, and the user loading on the performance of the proposed receiver are discussed
in the paper. The bit error rate (BER) performance of the proposed receiver is
evaluated by computer simulations. It has been found that the proposed receiver
provides a good BER performance, even at a large number of interfering users. At a
BER of 10 -3, the performance gain of the proposed receiver is about 2 dB over the
RAKE receiver with clipper decision and parallel interference cancellation in the
half loaded case (8 users ) and is much larger in the full loaded case (16 users).
Keywords: Downlink CDMA, Decision Functions, PIC, FDE-RZF, Zero Padding,
Channel Estimation.

INTRODUCTION

Recently, single-carrier transmission with

frequency domain equalization (FDE) has attracted
much attention for its excellent performance even in
strong frequency selective channels. In practice, the
number of fingers in the RAKE receiver is limited
because of hardware complexity. The use of an FDE
can alleviate the complexity problem of the RAKE
receiver arising from too many paths in a severe
frequency-selective channel. It has been shown that
the FDE can take the place of the conventional
RAKE receiver with much improved BER
performance for DS/CDMA signal reception over a
severe frequency-selective channel [1-5]. This gives
the CDMA with FDE the power to compete with
multi carrier CDMA (MC-CDMA) in fourth
generation systems. The performance of CDMA is
mainly limited by the interference from other users,
which is called the MAI. Therefore, PIC has to be
applied in CDMA receivers [6-13].
PIC has gained a considerable attention for its
potential ability to increase system capacity and its
simplicity. However, the conventional PIC often

Volume 2 Number 4

Page 75

suffers from the error propagation phenomena that

contributes to progressively enhanced interference
[12]. There are several algorithms for interference
cancellation in CDMA systems [ 6-13 ]. Most of
these algorithms are designed for the uplink. For
uplink interference cancellation, it is assumed that the
base station knows all the spreading codes of the
active users. For downlink CDMA, the receiver
knows only the spreading code of the desired user.
As a result, PIC has been assumed to be applicable at
the base station, and not at the mobile terminal where
only the information stream is to be decoded and the
spreading codes of the interfering users are unknown.
However, in last years, many algorithms have been
proposed for the estimation of the codes of
the
interfering users and PIC has been applied for
downlink CDMA [7]. The main target of this paper is
to analyze the performance of PIC for downlink
CDMA with different decision functions and hence to
develop an efficient receiver based on FDE and PIC
that is suitable for downlink CDMA.
The remainder of this paper is organized as
follows: in section 2, the system model of downlink
CDMA is presented. In Section 3, the concept of PIC

www.ubicc.org

Ubiquitous Computing and Communication Journal

is discussed. Section 4 deals with the different

possible decision functions. FDE for downlink ZPCDMA is presented in section 5. In Section 6, the
proposed FDE-RZF-CPIC algorithm is presented.
Channel estimation is discussed in section 7. The
relative performance of the proposed receiver is
compared to some existing approaches in section 8.
Finally, Section 9, gives the concluding remarks.
Notations: The symbols (.)H, (.)T, and (.)-1 designate
complex conjugate transposition of a matrix,
transposition of a matrix, and inverse of a matrix,
respectively. Vectors and matrices are represented in
boldface.
2

SYSTEM MODEL

In downlink CDMA, the channel is common

with frequency selective fading. We assume that the
channel parameters vary slowly with time, so that for
sufficiently short intervals the channel is
approximately a linear time-invariant system. The
baseband channel response can then be expressed by
Dirac
delta
functions
as
follows
[9]:
h (t ) =

h
W

(t w )

(1)

where hw , and w are the complex fading and

propagation delay of the w-th path, and W is the
number of multipath components of the channel
impulse response. In this paper, we assume block
fading, where the path gains stay constant over one
block duration. The received signal at the mobile can
be written as[9]:

r(t) = hw AKbk (l)c(t w )sk (t lTs w ) + n(t)

L KW

(2)

l k w

where Ak is the amplitude, bk(l) {-1,1} is the lth

bit, sk (t) is the spreading code of user k, and c(t) is
the scrambling code. In matrix form, the received
signal in Eq. (2) can be written as follows [6]:
r = H .d
0

+ H .d m 1 + n m
1

(3)

Where d m is the mth block of the transmitted signal,

rnoise
is the
vector,
while nmofis 2.the
withreceived
zero mean
and variance
H additive
is the
n
0 channel
(NL)(NL) matrix describing a multipath
having impulse response h(t) of length W. L is the
number of symbols for each user. N=TS/Tc is the
number of chips per bit (spreading factor). TS is the
symbol period. Tc is the chip period. K is the number
of active users.
We can observe that inter-block interference
(IBI) would disappear from Eq. (3) if the last W-1

Volume 2 Number 4

Page 76

elements of vectors d m and d m 1 were equal, i.e., if

a zero padding (or cyclic prefix) process is used. The
length of the zero padding must be greater than W. In
this paper we will use a zero padding method as in
[1]. So, Eq.(3) can then be written as:
r = Hd + n

(4)

Where H=H0+H1 has now a circulant structure.. H

can be written as:

h[0]
.

.
H = h[W1]
0
.

0 . 0 h[W1] . . h[0]
h[0] .
. .
.
.
.
. h[W1]

0
.

.
.
.

0 h[W1] . .

(5)

h[0]

The vector d can be represented as:

d = CSb

(6)

Where C is an (NL)(NL) scrambling code

matrix. S is an (NL)(KL) orthogonal code matrix,
and H 1d m1 is the inter block interference (IBI)
term [6]. H0 and H1 can be written as [6]:
. . . . . 0
0
h[0]
.

h
[
0
]

.
H0 = .

h[W 1]
.

.
.
.

0
. . h[W 1] . . . h[0]

, and
0

H1 = .
.

.
0

h[W 1]

.
.
.
.

h[0]

h[W 1]

(7)

The structures of the individual components in

Eq. (6 ) are given bellow [14]:
S = diag[ S S......S ]

(8)

www.ubicc.org

Ubiquitous Computing and Communication Journal

S = [s 1 s 2 ........ s K ]

s = [s
k

(9)

(0), sk (1),..........., sk ( N 1)]

(10)

b = [b (1)b (2).......b ( L)]

(11)

C = diag[c (1), c (2),......, c ( N L)]

(12)

b(l ) = [b1 (l ), b2 (l ),........., bK (l )]

In this paper, we suppose a perfect estimation of

the downlink interfering codes. So, we can write the
received signal as follows:
(14)

Where Sd is an ((NL)L) matrix consisting of the

spreading code of the desired user, U is an
(NL)((K-1)L) matrix consisting of the spreading
codes of the interfering users, bd is an L1 vector
consisting of the desired symbols, and bint is a ((K1)L)1 vector consisting of the interfering symbols.
In the zero padding method, NZP zeros will be
added to the end of NF-NZP symbols to build a block
of NF symbols before transmission. At the receiver,
the FFT detection will be performed on the padded
data block, the detected zeros at the end of this data
block will then be discarded after despreading. The
zero padding process is illustrated in Fig. (1).
3

PIC FOR DOWNLINK ZP-CDMA

Parallel interference cancellation for CDMA

systems has attracted much interest, due to its
structured
architecture
that
facilitates
easy
implementation. It was first introduced in 1990 [15].
Such multistage PIC methods attempt to cancel MAI
based on tentative decisions. The idea of PIC is to
estimate the multiple access and multipath induced
interferences and then to subtract the interference
estimate. The circulant matrix H can be efficiently
diagonalized by the fast Fourier transform (IFFT)
and inverse fast Fourier transform (FFT). Let -1 and
denote the FFT matrix and the IFFT matrix,
respectively.
A circulant matrix A can be written as (see appendix
1) :
1
(15)
A =
where is the FFT of the circulant sequence of A.
The implementation of the RAKE receiver with PIC

Volume 2 Number 4

d RAKE = (

Page 77

R)

(16)

Where is a diagonal matrix containing the FFT of

the circulant sequence of H, and RT is the FFT of r.
2-Estimate all interferences as follows:
T H
b int = U C d

(13)

Where SK is the signature sequence (code), and b is a

vector consisting of the users amplitudes and the
transmitted bits, and bK(l) is the lth bit.

r = HCS d b d + HCUbint + n

in frequency domain can be summarized as follows

[6]:
1- Apply the RAKE receiver on the received signal as
follows [14]:

(17)

RAKE

3- Discard the detected zero symbols at the end of

each block to produce b int , then take the decision as

follows:

L
~ = f {b }
dec
int
b int

(18)

where fdec(.) is the tentative decision function.

4- Add zeros for padding , then regenerate the MAI
as follows:
r MAI = H CU b~ ' int
(19)

~'

where b int is the zero padded version of b int .

5- Use PIC to cancel the effects of interference on the
received signal to obtain an interference free signal:
z = r rMAI

(20)

6- Apply the RAKE receiver on the vector z as

follows:
d RAKE = (

z)

(21)

7- Descramble and despread the obtained data.

8- Finally, discard the detected zero symbols and
perform hard decision.
Due to error propagation, PIC
with hard
decision may perform worse than PIC with linear or
soft decision functions. On the other hand, harddecision interference cancellation can completely
cancel interference when the hard decisions made are
correct [12].
4

DECISION FUNCTIONS

The performance of PIC depends on the decision

function used in the interference cancellation
iterations, e.g., hard, soft, null zone, unit clipper, and
hybrid decision functions [12].
The general model for the decision function is:

y = f dec ( x)

(22)

The following, decision functions can be used:

The hard limiter [12]:

www.ubicc.org

Ubiquitous Computing and Communication Journal

16 zeros

1 block (256 chips)

1 slot data (2560 chips )

Block- by block detection

FFT detect
+despreading

FFT detect
+despreading

FFT detect
+despreading

Figure 1: CDMA transmission for FDE detector using zero padding

1 x 0

y = fdec(x) =
1 x < 0

y
x

(23)

x >1
1,

y = fdec(x) = x, x [1,1]

x < 1
1,

Hard Limiter

It makes a hard decision for one of the two

possible symbols.

It makes a soft bit decision when the soft bit

estimate lies inside the interval [ -1 , 1 ] to avoid
propagation error , and makes a hard decision when
the soft bit estimate lies outside the interval [-1 , 1] to
avoid the noise magnification [10].

The null zone function [11]:

y

1, x > cn

y = f dec(x) = 0, x [cn ,cn ]

1, x < cn

1
-cn

-1

(26)

(24)

cn

-1

FREQUENCY DOMAIN EQUALIZER FOR

DOWNLINK ZP-CDMA.

Null Zone

It makes a hard decision when the soft bit

estimate lies outside the interval [-cn,cn], and sets the
decision result to zero when the soft bit estimate lies
inside the interval [-cn , cn]. Where cn is the null
zone decision threshold (0 cn 1)
[11].
The linear decision function:

y = f dec ( x) = x

(25)

It offers analytical access to the PIC performance,

but performs worse than other decision functions.
The unit clipper decision function [11]:

d ZF = ( H

H )

(27)

To encounter the problem of noise enhancement

in ZF equalizers, a new regularization term is added
into Eq. (27) to yields:

The application of FDE techniques makes single

carrier modulation a potentially valuable alternative
to OFDM, especially in regard to its robustness to RF
implementation impairments. Linear ZF based chip
level equalization has been one of the most popular
equalizers for single user downlink CDMA [16].
Because of the noise enhancement in the ZF
equalizers, we propose the application of the
regularized zero forcing equalizer.
The time domain ZF estimation of d is given by [16]:

-1
1
-1

Unit Clipper

Volume 2 Number 4

Page 78

www.ubicc.org

Ubiquitous Computing and Communication Journal

H
1 H
d RZF = (H H + ) H r

= M 1H H r = Gr

(28)

Where is a regularization parameter .

The solution of Eq. (28) requires the inversion of
the matrix M which has dimensions of
(NL)(NL). This inversion process is impractical
in real time. Thus, a simplification is required for this
inversion process.
The equalizer matrix G can then be easily calculated
as follows:
G =

1
M

(29)

RAKE receiver with CPIC to estimate, regenerate,

and cancel all the interfering users. Then the FDERZF equalizer is used to reduce the ISI effect and to
provide better estimate of desired user's data. In this
section, a specific data detection algorithm for
downlink CDMA is derived which is based on FDERZF equalizer and CPIC. The proposed FDE-RZFCPIC system model is depicted in Fig. (2).
The FDE-RZF-CPIC algorithm can be summarized
as follows:
1- Estimate all interferences as follows:
b int = U T C H H .R T

(33)

where:

1
M

= [ + I]
H

2- Discard the detected zero symbols at the end of

(30)

each block to produce b int , then take the decision as

follows:

The FDE-RZF algorithm can be summarized as

follows:
1- Apply the FDE-RZF on the received signal as
follows:
d FDE RZF = (

1
M

RT)

(31)

3- Then, a better estimation of the symbol of

interest can be obtained as follows:
T
H
b d = S d C d FDE RZF

(34)

where fdec(.) is a decision function that transforms

the soft estimate into a unit clipper decision.
3- Add zeros for padding , then regenerate the MAI
as follows:
(35)
r MAI = H CU ~b ' int

~'

where b

int

is a zero padded version of b int .

(32)

4- Finally, discard the detected zero symbols at the

end of each block, and then use the decision function.
A major advantage of this equalization method is
its low computational complexity. The price to be
paid is a reduction of the data rate caused by
insertion of zero padding or cyclic prefix.

4-Use PIC to cancel the effects of interference on

the received signal to get an interference free
signal:
Z = r rMAI

The complexity of a P-point FFT is of the order of

Plog2P. The FDE provides a complexity of (P log2
P ) which shows a significant reduction as compared
to that of the direct inversion of a matrix of
dimensions PP that has a complexity of the order of
(P3) [1]. The FDE has also less complexity than
that of the RAKE receiver which has a complexity
of (P2) [1].
6 FDE-RZF WITH CLIPPER PIC
This section gives the proposed receiver which is
used to improve the performance of the PIC for
downlink CDMA. The proposed receiver uses the

Page 79

(36)

5- Apply the FDE-RZF to the signal vector Z as

follows:

d FDECPIC

5.1 Complexity

Volume 2 Number 4

L
~
b int = f dec {b int }

= ( 1M . H . 1 (Z))

(37)

6- Then, a better estimation of the symbols of

interest can be obtained as follows:
T
H
b d = S d C d FDE CPIC

(38)

7- Finally, discard the detected zero symbols at the

end of each block, and then use the decision function.
The performance of FDE-RZF-CPIC depends
heavily on the channel estimate, not only in the
detection step, but also in the interference regeneration
step. It is more efficient when the system is heavily
loaded.

www.ubicc.org

Ubiquitous Computing and Communication Journal

downlink synchronous ZP-CDMA system, in which

each user transmits BPSK symbols. These symbols
are spread. After spreading, the resulting sum signal
is scrambled using a complex scrambling sequence.
The propagation channel is taken to be chip-spaced
with delay spread of 3Tc. For all simulations, we
take N=16, and the block size is NF=256 chips with
16 zeros (NZP=16) at the end of each block as shown
in Fig. (1). All users are assigned the same power.

7 CHANNEL ESTIMATION
In this section, we consider the channel estimation
method which depends on the pilot signal. When the
pilot sequence is transmitted, the received signal in
Eq. (4) is expressed as:
r = Dh + n
(39)
where the complex channel impulse response h is
expressed as:
FDE-RZF

Implementation
of

( +I)
H

Channel Estimation
Received
signal

-1

Descrambling,
Despreading,
And
remove ZP

IFFT

Hard
decision

desired
user's
data

RAKE

FFT

Genera
tion of

IFFT

Descrambling
Despreading,
And
remove ZP

Unit
Clipper
decision

2
.
.
.

Interference
Regeneration

Figure 2: The structure of FDE-RZF-CPIC for downlink ZP-CDMA system.

h = [h1 , h2 ,.........., hW ]

(40)

D is the circulant pilot sequence matrix.

The MMSE channel estimates are found by
minimizing the following squared error quantity:

h = arg min r Dh

Figure 3 compares the performance of the FDERZF and that of the TDE and the RAKE receiver
for 8 users . It is clear that the equalization in the
time domain is identical to that in the frequency
domain. The only difference is in the method of
implementation.
Both equalizers have better
performance than that of the RAKE receiver only.

(41)

SF=16,K=8, =1

10

RAKE

Assuming white gaussian noise, the ZF solution is

given by:
(42)

However, using zero forcing channel estimation,

the channel estimation accuracy significantly
degrades due to the noise enhancement.

AVERAGE BER

h ZF = (DH D) 1 D Hr

Zero padding

TDE-RZF Zero Padding

FDE-RZF Zero padding

10

-1

-2

10

8 SIMULATION RESULTS
Several simulation experiments are carried out to
test the performance of the proposed FDE-RZFCPIC algorithm and compare it to other algorithms.
The simulation environment is based on the

-3

10

10

15

SNR

Figure 3: Performance of FDE-RZF, TDE-RZF and

RAKE receiver Vs the SNR .

Volume 2 Number 4

Page 80

www.ubicc.org

Ubiquitous Computing and Communication Journal

The performance of the RAKE receiver with PIC

and different decision functions is studied. Figures 4,
and 5 illustrate the average BER versus the
threshold of the null zone decision function (cn) at
different SNR values and different number of users.

single-user RAKE receiver is very poor, even for

high SNR values. Parallel interference cancellation
improves the performance significantly. Better
performance can be obtained with the clipper
decision function. Linear decision performance is
worse than that of the single user RAKE receiver for
the heavily loaded case (K=16). This is justified by
the fact that PIC with linear decision is limited by
noise enhancement.

SF=16,K=8

SF=16,K=8

10

10
NULL ZONE DECISION

-1

-1

SNR=[ 6 9 12 15] db

10
AVERAGE BER

AVERAGE BER

10

-2

10

-3

10

10

-2

RAKE

10

-3

Null zone decision

unit clipper decision
hard decision

-4

10

0.2

0.4

0.6

0.8

10

THRESHOLD

10

15

Figure 6: Performance of the RAKE receiver with PIC at

different decision functions Vs the SNR for the half loaded
case. cn=0.4.
SF=16,K=16

SF=16, K=16

SNR

Figure 4: Performance of the RAKE receiver with PIC Vs

null zone decision threshold (cn) at different SNRs .

10

sof t decision

-4

10
NULL ZONE DECISION

10

10

-1

SNR=[ 6 9 12 15] db
AVERAGE BER

AVERAGE BER

10

-2

-1

10

-2

10

RAKE

-3

Null zone decision

unit clipper decision
hard decision

10

-4

sof t decision

-3

0.2

0.4

0.6

0.8

10

10

15

Figure 7: Performance of the RAKE receiver with PIC

at different decision threshold functions Vs the SNR
for a full loaded case. cn=0.4.

Figure 5: Performance of RAKE with PIC Vs null zone

decision threshold (cn) at different SNR .

The optimal performance is obtained when

cn=0.4. Figures 4, and 5 show that cn is nonsensitive to SNR-changes and to system-load
changes.
The effect of the tentative decision function on
the performance of PIC for K=8 (half loaded), and
K=16 (Full loaded ) are studied and shown in Figs. 6,
and 7. As the number of users increases, the
performance deteriorates. The performance of a

Volume 2 Number 4

SNR

THRESHOLD

Page 81

The performance of the proposed FDE-RZFCPIC is compared to that of the RAKE receiver,
FDE-RZF equalizer, the RAKE receiver with hard
decision PIC, and the RAKE receiver
with unit
clipper decision PIC.
The effect of the regularization parameter on the
performance of FDE-RZF-CPIC is examined in two
experiments and shown in Figs. 8 and 9. The optimal

www.ubicc.org

Ubiquitous Computing and Communication Journal

SF=16,K=16

10

-1

10
AVER AG BER

value of the regularization parameter is equal to 1.

This value is neither sensitive to SNR-changes nor
to system-load changes.
The effect of the choice of the tentative decision
function on the performance of the proposed
receiver is studied ( Fig. 10 ). The performance with
clipper decision outperforms the performance with
all other decision functions .

-2

10

SF=16,K=8

RAKE

10

-3

10

FDE-RZF w ith Null zone decision PIC

FDE-RZF w ith CPIC

FDE-RZF-CPIC

FDE-RZF w ith hard decision PIC

-1

AVERAGE BER

10

SNR=[ 6 12 ] db

FDE-RZF w ith linear decision PIC

-4

10

10

15

SNR
-2

10

Figure 10: Performance of the proposed receiver

at different decision Functions Vs the SNR.

-3

10

Figures 11, and 12 show the performance of five

reception schemes as a function of SNR of each user
for 8 and 16 users, respectively. From Figs. 11, and
12, it can be observed that there is a clear
improvement achieved by FDE-RZFCPIC scheme
over other
reception schemes. In Fig. 12, BER
performances of all receivers are worse than the
performances in Fig. 11 because of the increment in
the number of users. The FDE-RZF-CPIC scheme
improves the performance significantly, without
saturation of the performance for high SNRs like the
RAKE receiver. For the heavily loaded case (Fig.
12), the performance of FDE-RZF equalizer is
greater than that of the RAKE with PIC scheme. This
can be explained by the fact that at heavily loads, the
RAKE receiver sees too much interference, which
makes its decisions about interfering users unreliable.

-4

10

10

-2

10

Regularization Param eter ( )

Figure 8: Performance of FDE-RZF-CPIC scheme

Vs regularization parameter ( ) at different SNR
for a half loaded system. .
SF=16,K=16

10

FDE-RZF-CPIC

AVERAGE BER

10

-1

SNR=[ 6 12 ] db

-2

10

SF=16,K=8, =1

10

10

-3

-1

10
10

10

-3

10

-2

-1

10

10

Regularization Param eter ( )

10

Figure
9: Performance
scheme
( FDE-RZF-CPIC
), at different SNR
for a Vs
full
regularization
parameter of
loaded system..

AVERAGE BER

-4
-2

10

-3

10

RAKE

10

FDE-RZF

-4

hard decision rake+pic

unit clipper decision rake+pic
FDE-RZF-CPIC

10

15

SNR

Figure 11: Performance of different reception schemes Vs

the SNR for a half loaded system.

Volume 2 Number 4

Page 82

www.ubicc.org

Ubiquitous Computing and Communication Journal

10

users increases.

SF=16,K=16, =1

RA KE

SF=16

10

FDE-RZF
hard decision rake+pic

FDE-RZF-CPIC

unit clipper decision rake+pic

-1

FDE-RZF-CPIC

-1

10
AVERAGE BER

AVERAGE BER

10

-2

10

-2

10

-3

10

-3

10

15

10

SNR

10

14

12

Num ber of Cancelled Us ers

Figure 12: Performance of different reception schemes Vs

the SNR for a full loaded system.

Figure
14:canceled
Performance
FDE-RZF-CPIC
scheme Vs
number of
users. of
=1,
and SNR =12 dB.

The effect of user loading on the performance of

the FDE-RZF-CPIC scheme is studied and presented
in Fig. 13. The BER of all receivers degrade with
increasing the number users. In this case, the BER
performance of FDE-RZF-CPIC scheme degrades a
little bit with increasing the number of users, but it is
still better than the other schemes. This observation
may be due to the MAI. The MAI when the number
of users is large should be greater than the case when
the number of users is low. Even after interference
cancellation, some residual MAI still exists.
Therefore, the performance loss may be attributed to
the residual MAI.

The effect of channel estimation accuracy on the

performance of the FDE-RZF-CPIC scheme for K=8
are studied and shown in Figs. 15, and 16. The
performance of FDE-RZF-CPIC scheme with ZF
channel estimation shows a loss of 1 dB at BER of
10-2 when compared with the case of perfect channel
knowledge. Because the noise enhancement in the
ZF channel estimation. LMMSE channel estimation
gives better performance.

rake+pic chann. know n

rake+pic chann. estimation
FDE-RZF+CPIC chann. know n

SF=16, =1

10

FDE-RZF+CPIC chann. estimation

-1

10
AVERAG BER

RAKE
FDE-RZF
hard decision rake+pic

-1

10
AVERAG BER

SF=16,K=8, LMMSE Channel estimation

10

unit clipper decision rake+pic

FDE-RZF w ith CPIC

-2

10
-2

10

-3

10

-3

10

10

12

SNR

Figure 15: Performance of Interference Cancellation

Vs the SNR for exact and LMMSE channel estimate.

-4

10

8
10
Number of Users ,K

12

14

Figure 13: Performance of different reception schemes Vs

the number of active users. and SNR =12 dB.

Figure 14 depicts the average BER performance

as a function of the number of cancelled users, at a
fixed SNR per user of SNR=12 dB. This graph
shows that the performance of the FDE-RZF-CPIC
scheme improves when the number of cancelled

Volume 2 Number 4

Page 83

www.ubicc.org

Ubiquitous Computing and Communication Journal

(A.2)
where each raw is a circular shift of the raw above
and the first raw is a circular shift of the last raw.
The primary difference between the matrices Q and

SF=16,K=8, zero padding ,ZF Channel estimation

10

-1

AVERAG BER

10

Q c is that they differ only by elements added in the

upper right and lower left parts to produce the cyclic
structure in the raws. If the matrix size S is large
and the number of non zero elements on the main
diagonals compared to the number of zero elements
is small (i.e the matrix is sparse), the number of
elements added to the upper right and lower left parts
does not affect the matrix, because they are small in
proportion to the main diagonal elements. It can be
shown from the eigen value distribution of both

-2

10

-3

10

rake+pic chann. known

rake+pic chann. estimation
FDE-RZ F+CPIC chann. known
FDE-RZ F+CPIC chann. estimation

-4

10

10

12

matrices that Q
equivalent.

SNR

Figure 16: Performance of Interference Cancellation Vs

the SNR for exact and ZF channel estimate.

9 CONCLUSION
The paper presents an efficient FDE-RZF-CPIC
receiver for downlink CDMA. This receiver is
implemented using frequency domain approximations rather than the time domain implementation to
reduce complexity. The comparison studies show that
the proposed receiver outperforms several traditional
receivers for different loading cases. The sensitivity
of the proposed receiver is also studied for different
decision functions and different channel estimation
methods. The obtained results indicate that the
proposed receiver performance is robust for the
different channel estimation methods.
APPENDEX 1 Toeplitz to circulant approximation
Let Q be an SXS Toeplitz matrix of the following
form:
0
q(0) " q(l )

# %

Q = q(k )

q(l )
%

%
#
q(k ) " q(0)
%

(A.1)

It can be approximated by an SXS circulant matrix

Q c defined as [17,18]:
q(0) " " q (l )
# %
#
%
q(k )
%
c
Q =
0
%

# % %
%
q (l ) # %
#
%
#
%

q ( 1) " q (l ) #

Volume 2 Number 4

"

q (k ) "
"

"

%
%

%
%

%
%
%

%
0

q(k )

"

"

q(k )
"

q (l )

#
#
q (1)

c
and Q are asymptotically

c
It is known that an SXS circulant matrix Q is
diagonalized by [1]:
1 c
(A.3)
= Q
where is an SXS diagonal matrix whose
c
elements (s, s) are the eigenvalues of Q and
where is an SXS unitary matrix of eigen vectors
c
of Q . Thus we have:
= = I
The elements (s1, s2 ) of
[17,18]:
j2s s
(s1, s2 ) = exp 1 2
S
*t

*t

(A.4)
are given by

(A.5)

2
= 1
for s1, s2 = 0,1,........, S 1and j
The eigen values (s, s) can be called (s) . For
these eigen values, the following relation holds
[17,18]:
k
j 2ms
(s) = q(0) + q(m) exp
S
m=1
(A.6)
1
j 2ms
+ q(m) exp
S
m=l
s = 0,1,........, S 1
c
Because of the cyclic nature of Q ,we define:
(A.7)
q(S m) = q(m)
and thus Eq.(A.6) can be written in the form [17,18]:
S 1
j 2ms
(s) = q(m) exp
(A.8)

m=0
S

for s = 0,1,........, S 1
Thus the circulant matrix can be simply diagonalized
by computing the DFT of the cyclic sequence
q(0), q(1),......., q(S 1) .

q (0)

Page 84

www.ubicc.org

Ubiquitous Computing and Communication Journal

10 REFERENCES
[1] I. Martoyo, T. Wesis, F. Capar, and F.Jondral,
Low complexity cdma downlink
receiver
based on frequency domain equalization, in
Proc. IEEE Vech. Tech.. Conf., pp. 987-991,
Oct. 2003.
[2] J. Pan, P. De, and A Zeira, Low complexity
Data Detection Using Fast Fourier Transform
Decomposition of Channel Correlation Matrix,
IEEE Global Telecom. Conference, vol. 2 , pp.
1322-1326, Nov. 2001.
[3] D. Falconer, S. L. Ariyavisitakul, A. BenyaminSeeyar, and B. Edison, Frequency domain
equalization for single-carrier broadband
wireless systems, IEEE Mag. Commun., Vol.
40, no. 4, pp. 58-66, Apr. 2002.
[4] F. petre, G. Lues, L. Deneire, and M. Moonen,
Downlink frequency domain chip equalization
for single-carrier block transmission ds-cdma
with known symbol
padding, in Proc.
GLOBCOM, pp. 453- 457, Nov. 2002 .
[5] K.L. Baum, T.A. Thomas, F.W Vook, and V.
Nangia, Cyclic-prefix CDMA: an Improved
Transmission Method for Broadband DSCDMA Cellular Systems,
IEEE Wireless
Comm. And
Networking Conference, vol. 1,
pp.17-21, Mar. 2002.
[6] B. Mouhouche, K. Abed-Meraim, N. Ibrahim,
and P.
Loubaton,
Combined
MMSE
equalization and blind parallel interference
cancellation for downlink multirate CDMA
communications,
IEEE 5th
Workshop on
Signal Processing Advances in Wireless
Communications, pp. 492496, 11-14 July
2004.
[7] B. Mouhouche, K. Abed-Meraim, and S. Burykh,
Spreading code detection and blind interference
cancellation for DS/CDMA downlink, IEEEISS
STA 04, Sydney, Australia, pp. 774-778, 2004.
[8] Z. Gao, Q. Wu, and J. Wang, Combination of
LMMSE equalization and multi-path interference cancellation in WCDMA receivers IEEE
ICMMT 4th International Conference, pp. 838
841, 18-21 Aug. 2004.
[9] M.F. Madkour, S. C. Gupta, and Y. .E. Wang,
Successive Interference Cancellation Algorithms for Downlink W-CDMA Communications,
IEEE Trans. Wireless Comm., vol. 1, N. 1 ,
January 2002.
[10] L. B. Nelson, and H. V. poor, Iterative
multiuser receivers for CDMA channels: An
EM- based approach, IEEE Trans. Commun.,
Vol. 44, pp. 1700-1710, Dec. 1996.
[11] A. L. C. Hui, and K. B. Letaief, Multiuser
asynchronous DS/CDMA detectors in multipath
fading links, IEEE Trans. Commun., vol. 46,
pp. 384391, Mar.1998.

Volume 2 Number 4

Page 85

[12] W.Zha, S., and D. Blostein , Soft-decision

multistage multiuser interference cancellation,
IEEE Trans. Techn., Vol. 52, No. 2, pp. 380389, 2003.
[13] A. Klein, G. K. Kaleh, and P. W. Baier, Zero
Forcing and Minimum Mean-Square-Error
Equalization for Multi-User Detection in Code
Division Multiple-Access Channels, IEEE
Trans. Vehic. Tech., vol. 45, no. 2, pp. 276-87
May 1996.
[14] S. Werner, and J. Lilleberg, Downlink channel
decorrelation in CDMA systems with long
codes, IEEE 49th Vehicular Technology
Conference, vol. 2, pp. 16141617, 16-20, 1999.
[15] M. Varanasi, and B. Aazhang, Multistage
detection in Asynchronous Code-Division
Multiple Access Communications, IEEE Trans.
Commun. , 38(4), pp. 509-519, 1990.
[16] A. Klein, Data detection algorithms specially
designed for the downlink of CDMA mobile
radio systems, IEEE 47th Vehicular Technology
Conference, vol. 1, pp. 203207, 4-7 May 1997.
[17] H.C. Anderws and B.R. Hunt, Digital Image
Restoration. Englewood Cliffs, NJ: PrenticeHall, 1977.
[18] Jan Biemond , Jelle Rieske and Jan J.
Gerbrands, A Fast Kalman Filter For Images
Degraded By Both Blur And Noise,
IEEE
Trans. Acoustics, Speech and Signal Processing,
Vol. ASSP-31, No.5, pp.1248-1256, Oct. 1983.

www.ubicc.org